Quantification of Diatom Gene Expression in the Sea by Selecting Uniformly Transcribed mRNA as the Basis for Normalization

Abstract

To quantify gene expressions by quantitative reverse transcription-PCR (Q-RT-PCR) in natural diatom assemblages, it is necessary to seek a biomass reference specific to the target species. Two housekeeping genes, TBP (encoding the TATA box-binding protein) and EFL (encoding the translation elongation factor-like protein), were evaluated as candidates for reference genes in Q-RT-PCR assays. Transcript levels of TBP and EFL were relatively stable under various test conditions including growth stages, light-dark cycle phases, and nutrient stresses in Skeletonema costatum and Chaetoceros affinis, and TBP expression was more stable than that of EFL. Next, the sequence diversity of diatom assemblages was evaluated by obtaining 32 EFL and 29 TBP homologous gene fragments from the East China Sea (ECS). Based on sequence alignments, EFL and TBP primer sets were designed for Chaetoceros and Skeletonema groups in the ECS. An evaluation of primer specificity and PCR efficiency indicated that the EFL primer sets performed better. To demonstrate the applicability of EFL primer sets in the ECS, they were employed to measure mRNA levels of the FcpB (fucoxanthin-chlorophyll protein) gene in diatoms. The results correctly revealed prominent diel variations in FcpB expression and confirmed EFL as a good reference gene.

INTRODUCTION

Marine phytoplankton consists of a group of genetically diverse microorganisms that contribute more than 46% of global net primary production (9). In marine ecosystems, phytoplankton not only produces organic carbon for the entire food web but also controls pathways of carbon flow via the formation of dominant species. The ability to evaluate the status of phytoplankton health in natural environments is thus important to understanding how phytoplankton tolerates environmental stresses and copes with interspecific competition. With advances in molecular biological techniques, molecular markers were proposed as powerful indicators to detect the physiological status of phytoplankton (21, 29). For example, mRNA levels of the nitrate transporter gene, Nrt2, are a good molecular marker of diatom nitrogen deficiency (17, 34). mRNA levels are low in the presence of ammonium but increase 1,000-fold when ambient nitrogenous nutrients are exhausted (18). The advantages of this approach include a unified detection method for all environmental stresses, high sensitivity, and species or genus specificity (19).

Among various mRNA detection and quantification methods, quantitative reverse transcription-PCR (Q-RT-PCR) has been the most popular in recent years due to its accuracy and sensitivity (2, 36). However, applying this method to detect mRNA abundances in natural environments faces the challenge of samples containing mixed species (33). In total RNA extracted from an environmental sample, primers are usually designed to quantify the mRNA of a single target gene in a single target species. The purpose is to avoid interspecific variations in target mRNA production, which complicate data interpretation. In a given sample, this measured mRNA level is the product of the cell abundance and cellular content of the target mRNA in a single species. To make a fair comparison between samples, a reference that can faithfully represent the biomass of the target species is required to normalize the target mRNA levels measured by Q-RT-PCR.

In laboratory cultures, housekeeping genes are widely used as biomass references when quantifying mRNA levels of genes under study (3, 13). These housekeeping genes are typically involved in basic cellular processes with a constitutive expression at mRNA levels that fluctuate minimally. Supposedly, the same approach is applicable to natural samples, as long as the primers for the target and reference genes show good specificity to the species of interest. In this case, the transcript level of a housekeeping gene functions as an internal control and represents the biomass of the chosen species. In practice, the constancy of mRNA levels requires evaluation for species newly brought to our attention, since the performances of commonly recognized housekeeping genes vary considerably under different test conditions and in different organisms (6, 7). In studying phytoplankton nutrient stress in the ocean, species- and genus-specific primers were successfully used to detect mRNA levels of diatom nitrate transporter genes in the East China Sea (ECS) (19). A full evaluation of nitrogen stress in diatoms will require the selection of a suitable reference gene.

In diatoms, several housekeeping genes were proposed as candidates for a reference gene under a light-dark cycle. The TBP (TATA box-binding protein) gene encodes a subunit of the eukaryotic transcription factor TFIID and was proposed as a stable reference gene in various organisms (11, 25, 30, 31). It was also validated as one of the most stable reference genes for a diatom grown under a light-dark cycle (32). The EFL gene encodes a translation elongation factor-like protein that bears sequence similarity to the elongation factor 1α (EF-1α) protein. Homologs of the EFL gene were found in a wide variety of marine phytoplankton organisms, including diatoms, haptophytes, prasinophytes, and dinoflagellates (16, 20, 26). The expression of EFL was also stable under a light-dark cycle in a tested diatom (16). However, a reliable internal control in natural environments requires more than a minimal diel variation. Expression patterns under other growth conditions, such as nutrient stress, should be examined as well.

In this study, we evaluated the suitability of TBP and EFL as reference genes for Q-RT-PCR assays in natural assemblages of phytoplankton. We examined their transcript levels in two diatom strains including Skeletonema costatum and Chaetoceros affinis under different growth conditions and in different phases of a light-dark cycle. Next, TBP and EFL gene diversities were investigated in the ECS so that we could design species- and genus-specific primers. After their specificity and PCR efficiency were determined in environmental samples, EFL was identified as a good reference gene, and a very convincing pattern of EFL-normalized mRNA level of FcpB (gene expressing fucoxanthin-chlorophyll protein) was obtained for 2 dominant diatom groups, Skeletonema and Chaetoceros, in the ECS.

MATERIALS AND METHODS

Phytoplankton strains and maintenance.

Five diatom cultures, Chaetoceros affinis CCMP 160, Ditylum brightwellii CCMP 358, Phaeodactylum tricornutum CCMP 632, Thalassiosira oceanica CCMP 1003, and Thalassiosira pseudonana CCMP 1335, were obtained from the Provasoli-Guillard National Center for Marine Algae and Microbiota (West Boothbay Harbor, ME). The unialgal culture of Skeletonema costatum strain Kao was kindly provided by H.-M. Su of Tungkang Marine Laboratory, Pingtung, Taiwan (14), and Thalassiosira weissflogii was provided by E. Cosper (8). Additionally, the unialgal culture of Thalassionema sp. was established by micropipetting individual cells of natural samples from the ECS (J. Chang, H.-F. Wang, P.-S. Wu, L.-F. Chien, and C.-W. Liao, submitted for publication). All cultures were grown at 20°C in f/2-enriched seawater medium (12) under a 12:12-h light-dark photoperiod with a light intensity of 145 μE m⁻² s⁻¹.

Growth conditions for laboratory experiments.

Gene expressions under different nutrient conditions were evaluated in 3 types of media. Full-strength h/2 medium contained 500 μM ammonium as the sole nitrogen source (12). Low-ammonium (LN) medium was prepared by reducing the initial concentration of ammonium to 25 μM, and low-phosphate (LP) medium was prepared by reducing the initial phosphate concentration to 1.0 μM. Polycarbonate 2.8-liter Fernbach flasks were used as the culture vessels, and each contained 2 liters of culture media. During the experimental period, cells were harvested daily for cell counts, fluorescence measurements, and RNA extraction. Cell concentrations were determined with a Sedgwick-Rafter counting chamber (Hausser Scientific, Horsham, PA). The maximum photochemical efficiency of photosystem II (Fv/Fm) was measured on a fluorescence induction and relaxation (FIRe) fluorometer (Satlantic, Halifax, Canada) following procedures described in reference 24. Algal cells for RNA extraction were harvested by filtration using a 5-μm-pore-size Nuclepore polycarbonate membrane. Cells on the membrane were scraped off, resuspended in 0.7 ml RLT buffer (Qiagen, Valencia, CA) with 10 μl ml⁻¹ of β-mercaptoethanol (Sigma-Aldrich, St. Louis, MO), and then stored at −80°C until further processing. Cells used for diel-variation studies were grown in f/2 medium under a 12:12-h light-dark photoperiod. These batch cultures were grown to the late exponential phase, and cells were then harvested every 3 h during a 24-h period following the same sampling procedures as those described above.

Cruises and field sampling.

A research cruise was conducted in the southern ECS on 18 and 19 August 2009 on board the R/V Ocean Researcher II. Sampling was performed approximately every 3 h at a station located at 121.9°E, 25.5°N. Phytoplankton samples were collected with a 20-μm-mesh plankton net with a mouth diameter of 0.5 m. An oblique tow was performed from a 5-m depth to the sea surface for 10 min with the ship speed set to 1 knot. After the net tow, samples were removed from the receiving bottle (cod end) and filtered through a 200-μm mesh. Subsequently, plankton was collected with a 20-μm-mesh screen and transferred to cryotubes. Samples for DNA isolation were directly stored in liquid nitrogen. Samples for RNA isolation were resuspended in RLT buffer before being frozen in liquid nitrogen.

Genomic DNA and total RNA isolation.

Genomic DNA was extracted using the phenol-chloroform method with the addition of cetyltrimethylammonium bromide (5). Total RNA was isolated using an RNeasy Plant minikit (Qiagen) according to the manufacturer's instructions. The RNase-free DNase I set (Qiagen) was applied for on-column digestion of residual DNA during RNA purification. Concentrations of RNA and DNA were determined with a spectrophotometer (ND-1000; NanoDrop Technologies, Wilmington, DE) at wavelengths of 260 and 280 nm.

PCR amplification and DNA sequencing.

Degenerate primers for amplifying the TBP and EFL genes were designed from the conserved regions of homologous amino acid sequences in algae (Table l). To amplify TBP and EFL gene fragments from the ECS, 0.5 μg of genomic DNA extracted from environmental samples was used as the template in the PCR. PCR products were purified from a low-melting-point agarose gel and cloned into pGEM-T vectors (Promega, Madison, WI). DNA sequencing of the cloned fragments was performed using an ABI Prism 377A DNA sequencer with a PRISM Ready Reaction BigDye termination cycle sequencing kit (Applied Biosystems, Foster City, CA). Nucleotide sequences were analyzed by a BLAST search (http://www.ncbi.nlm.nih.gov/BLAST/).

Phylogenetic tree construction.

Nucleotide sequences of TBP and EFL were edited using DNAstar software (Lasergene, Madison, WI). Multiple alignments of sequences were performed using ClustalW (35), and aligned sequences were manually checked with the Bioedit sequence alignment editor (version 5.0.9; Department of Microbiology, North Carolina State University, Raleigh, NC). The resultant alignment files were used to construct phylogenetic trees with PHYLIP software (the PHYLogency Inference Package, http://evolution.genetics.washington.edu/phylip.html). Pairwise distances were calculated using the Kimura formula in the PROTDIST program, and neighbor-joining trees were generated with the Neighbor program. Bootstrap values were obtained with 1,000 bootstrap replicates to estimate the degree of confidence.

mRNA quantification for laboratory experiments.

To quantify mRNA, 0.5 μg of total RNA was reverse transcribed into cDNA using a high-capacity cDNA reverse transcription kit (Applied Biosystems). Quantitative PCRs were initiated by adding the cDNA into a mixture containing 1× SYBR green PCR master mix (Applied Biosystems) and 300 nM forward and reverse primers (Table 1). These primer sets targeting 18S rRNA, TBP, EFL, and FcpB were specifically designed for S. costatum and C. affinis (Primer Express software, Applied Biosystems). Quantitative PCRs were conducted using an ABI Prism 7500 Real-Time PCR system (Applied Biosystems). PCR settings were 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. The fluorescence intensity from the complex formed by SYBR green and the double-stranded PCR product was continuously monitored from cycles 1 to 40. The threshold cycle (C_T) at which the fluorescence intensity became higher than a preset threshold was used to compare transcript levels of individual genes.

Table 1.

Primers used in this study

Primer	Orientation	Nucleotidea sequence (5′→3′)	Target gene	Target species
Degenerate primers
DeflF1	Forward	AAGGARGARCGYGARCGTGG	EFL
DeflR2	Reverse	TCATYTTRCANGCRACCTTRGC	EFL
TbpF	Forward	TGCCGCAACACMGAATWYAAYCC	TBP
TbpR	Reverse	GGCACCMGTAATSACYAYCTTYCC	TBP
Primers for Q-PCR
D18SQF	Forward	GACTCAACACGGGAAAACTTACC	18S rRNA	Diatoms
D18SQR	Reverse	CACCAACTAAGAACGGCCATGC	18S rRNA	Diatoms
CaTBPQF	Forward	AAGGTTGGGTTTAAGACTGCTCC	TBP	C. affinis
CaTBPQR	Reverse	CGCATACACCAACCCCTCG	TBP	C. affinis
ScTBPQF	Forward	CTGCAGTAATTATGCGGCTGC	TBP	S. costatum
ScTBPQR	Reverse	CAGAATTGTGGGTGCTCTTCG	TBP	S. costatum
CaEFLQF	Forward	GGTCAGACCCGTCAGCACG	EFL	C. affinis
CaEFLQR	Reverse	AGGCATCCATCTTGTTGATTCC	EFL	C. affinis
ScEFLQF	Forward	TGGTCTTCTTCTTGTGCCTGC	EFL	S. costatum
ScEFLQR	Reverse	CGTGCTGACGGGTTTGTCC	EFL	S. costatum
CaFcpBQF	Forward	CGTAGAGATCAAGCACGGACG	FcpB	C. affinis
CaFcpBQR	Reverse	CGGTGTCGATGCACCTGG	FcpB	C. affinis
ScFcpBQF	Forward	CGAGCACGTTGGTGACTTCC	FcpB	S. costatum
ScFcpBQR	Reverse	GGCACGCTTCTGAAGCTTGG	FcpB	S. costatum
Group-specific primers
ChaTBPQF	Forward	AGGTTGGGTTCAAGACTGCTCC	TBP	Chaetoceros
ChaTBPQR	Reverse	CAACTCCGGCTCATAACTCGC	TBP	Chaetoceros
SkeTBPQF	Forward	GCTGCAGTAATTATGCGTCTGC	TBP	Skeletonema
SkeTBPQR	Reverse	CTTTTCGTCCCGGTGACAACC	TBP	Skeletonema
ChaEFLQF	Forward	CACCAAGGACGAGGTTGTTTCC	EFL	Chaetoceros
ChaEFLQR	Reverse	CTCCTGTACTACAACCATAGCGC	EFL	Chaetoceros
SkeEFLQF	Forward	CCCAAGAAGGCTGTTGAGGG	EFL	Skeletonema
SkeEFLQR	Reverse	GAAGACGGACGGGAAGAGTGG	EFL	Skeletonema

^aNucleotide single-letter codes: A, adenosine; C, cytosine; G, guanine; T, thymine; W, A or T; Y, C or T; R, A or G; M, A or C; K, G or T; S, G or C; N, A or T or C or G.

The absolute amount of target gene RNA (X) normalized to a unit of reference gene RNA (R) was calculated according to the absolute ratio method described in reference 4:

$$\log \left(\frac{X_0}{R_0}\right)=\log \left(\frac{M_R}{M_X}\right)+\frac{C_{\mathrm{T,X}}}{b_X}-\frac{C_{\mathrm{T,R}}}{b_R}$$ (1)

where X₀ and R₀ are the target and reference molecules in the original sample, M_R and M_X are the molecular weights of the target and reference amplicons, C_T,X and C_T,R are C_T values of X and R, and b_X and b_R are the standard curve slopes of X and R, respectively. Each Q-RT-PCR was performed in triplicate to enable calculation of a standard deviation for tube replicates:

$${\mathrm{Std}}_{\log (X_0/R_0)}=\sqrt{\frac{\mathrm{Var}\left(C_{\mathrm{T,X}}\right)}{b_X^2}+\frac{\mathrm{Var}\left(C_{\mathrm{T,R}}\right)}{b_R^2}}$$ (2)

where Var(C_T,X) and Var(C_T,R) are C_T variances of X and R, respectively. In time course experiments, log(X₀/R₀) was obtained for each sampling point, and the standard deviation for the sampling period was calculated according to the following equation:

$${\text{Std}}_{\text{log}(X_0/R_0)}=\sqrt{\frac{1}{m-1}{\displaystyle \sum _{i=1}^m{\left(y_i-\overline{y}\right)}^2}}$$ (3)

where y is log(X₀/R₀), and m is the number of sampling points. On a linear scale, the fold variation of tube replicates was calculated as 10 raised to the power of ${\mathrm{Std}}_{\log (X_0/R_0)}$ from equation 2. On the other hand, the fold variation for a sampling period was similarly calculated, but by using ${\mathrm{Std}}_{\log (X_0/R_0)}$ from equation 3. The specificity of the Q-RT-PCR products was confirmed by a melting-temperature analysis performed on the 7500 Real-Time PCR system (Applied Biosystems) from 65 to 95°C for 20 min and examined by electrophoresis on 3% agarose gels containing 0.5× Tris-boric acid-EDTA buffer.

Design and evaluation of group-specific primers.

Diatom group-specific primers suitable for applications in the ECS were designed in divergent regions of the TBP and EFL genes revealed by sequence alignments (Table 1). Detailed procedures and considerations are listed in reference 19. Each group-specific primer contained at least 4 mismatched nucleotides compared to homologous segments belonging to other diatom groups, and a single mismatch at the 3′ end was added to further reduce the between-group cross-reactions during PCR amplification.

In the cross-reactivity assay, genomic DNA (10 ng) of 8 cultivated diatoms was used to evaluate the specificities of group-specific primer sets (Table 1). For a primer set to be tested, a standard curve was constructed using serially diluted genomic DNA of the target diatom. Next, Q-PCR was performed on each of the 8 diatoms, using the same primer set, and an apparent amount of target gene in each species was obtained from the standard curve. The primer specificity was thus computed as the apparent amount in a particular species normalized to the apparent amount of the target species. On the other hand, the amplification efficiencies of these primer sets in their respective target species were evaluated using serially diluted genomic DNA (range, 0.01 to 10 ng) extracted from C. affinis and S. costatum. In addition, serially diluted environmental genomic DNA was used to evaluate PCR efficiencies in natural diatom assemblages.

Performance of EFL primers in the ECS.

To test if EFL is a good reference gene to reveal the cyclic diel expression pattern of FcpB, Q-PCR was performed using cDNA obtained from environmental samples collected during the ECS cruise. Four primer sets were used to individually detect mRNA of EFL and FcpB in 2 diatom groups: Skeletonema and Chaetoceros. Results of the Q-PCR were reported using both the absolute quantification method and the absolute ratio method. In the absolute quantification method, linearized plasmid DNA with precisely determined copy numbers of EFL and FcpB was used to construct standard curves. C_T values of environmental samples were converted to copy numbers, followed by normalization against the amount of total RNA used in the Q-RT-PCR. In the absolute ratio method, EFL-normalized FcpB mRNA levels were calculated according to equation 1. Standard curves required for the calculation were generated using serially diluted environmental genomic DNA (Table 2).

Table 2.

Properties of standard curves generated by various primer sets using unialgal-culture and environmental genomic DNAs as templates

Primer set	Culture genomic DNAa			ECS St. 9 genomic DNAd
	E%b	r2	Copy ratioc	E%	r2	Copy ratio
SkeEFL	98.0	0.99	1.00	85.7	0.95	1.00
SkeTBP	95.7	0.99	0.13	190	0.77
ScFcpB	93.4	0.99	27.8	91.3	0.99	18.4
ChaEFL	102	0.99	1.00	107	0.99	1.00
ChaTBP	104	0.99	2.00	1,814	0.14
CaFcpB	101	0.99	39.4	140	0.96	0.01

^aGenomic DNA from the target species was used as the template for standard curves.

^bAmplification efficiency (E%) = [10^(−1/slope) − 1] × 100%.

^cCopy ratio = 2^ΔCT, where ΔC_T is the average ΔC_T over a series of genomic DNA dilutions. ΔC_T was obtained from the difference between the C_T of the EFL gene and the C_T of another gene.

^dSt. 9 = 121.9°E, 25.5°N.

Nucleotide sequence accession numbers.

Sequences obtained in this study were deposited in GenBank under the accession numbers JQ762458 to JQ762524.

RESULTS

Constancy of EFL and TBP mRNA expressions.

When S. costatum was grown in nutrient-replete h/2 medium, the cell density exponentially increased for 5 days before entering the stationary phase. In contrast, cell populations in the low-ammonium and low-phosphate media grew more slowly during their exponential phases and entered the stationary phase on day 4 with lower maximum population densities, about one-third to one-half that in h/2 medium (Fig. 1A). In these low-nutrient cultures, values of the maximal photochemical efficiency of PSII (Fv/Fm) began to decline on day 4 (Fig. 1B). Jointly, these results indicate that growth was stressed by nitrogen or phosphorus after day 4.

Fig 1.

Time course of cell numbers, Fv/Fm, and gene transcript levels in batch cultures of Skeletonema costatum grown in nutrient-replete h/2 medium, low-ammonium (LN) medium, and low-phosphate (LP) medium. (A) Numbers of cells; (B) maximal photochemical efficiency of photosystem II (Fv/Fm); (C) elongation factor-like gene (EFL) transcript levels; and (D) TATA box-binding protein gene (TBP) transcript levels. The vertical scales on the right of panels C and D indicate the gene expression range of a death marker gene (ScDSP) and a nitrogen deficiency marker gene (TpNrt2.1). Error bars represent 1 standard deviation of replicate cell counts, Fv/Fm tube replicates, or Q-RT-PCR tube replicates. Data points without an error bar mean that the error bar is smaller than the symbol.

In h/2 medium during the experimental period, transcript levels of EFL were stable at 207 μmol (mol 18S rRNA)⁻¹ with a day-to-day variation of 1.47-fold on a linear scale. In the low-nutrient media, EFL mRNA began to increase from a basal level on day 3 and stabilized at a higher level at 1,767 μmol (mol 18S rRNA)⁻¹ on days 5 to 8 (Fig. 1C). Compared to genes of which the expression levels significantly changed between exponential and stationary phases, such as the death-specific protein gene, ScDSP (4), and the nitrate transporter gene, TpNrt2.1 (18), variations in the EFL mRNA level were rather small (Fig. 1C). Similar trends were observed for mRNA levels of TBP (Fig. 1D). We observed almost identical TBP mRNA levels in the exponential and stationary phases, and differences between low-nutrient and h/2 cultures were also small compared with those observed for EFL (Fig. 1D).

Diel variations of EFL and TBP expressions were examined in 2 diatom species grown under a light-dark cycle. 18S rRNA-normalized mRNA levels of both genes remained stable during the 24-h sampling period, with EFL mRNA being the more-abundant molecule (Fig. 2A and C). In the case of EFL in S. costatum, the daily averaged mRNA level was 2,410 μmol (mol 18S rRNA)⁻¹ with a variation of 1.59-fold during the 24-h period, which was almost as small as the variation of Q-RT-PCR tube replicates at 1.37-fold. In the 2 diatom species examined, daily averaged mRNA levels of housekeeping genes differed. With 18S rRNA as the normalization standard, EFL mRNA in S. costatum was about 25 times that in C. affinis (Fig. 2A and C). A similar result was obtained for TBP mRNA. Next, FcpB, a gene with a prominent diel variation in mRNA expression (28), was used to test the performances of 18S rRNA, EFL, and TBP as reference genes. Regardless of the reference gene used in the normalization, the FcpB transcript level always exhibited a dramatic decrease at the beginning of the dark period at 1930 h (Fig. 2B and D). Subsequently, the transcript level steadily increased with time and reached a peak during the light period at 1030 h. Within a diatom species, FcpB transcript levels normalized to TBP, EFL mRNA, and 18S rRNA fluctuated around different daily means (Fig. 2B and D). Due to the high abundance of 18S rRNA in cells, FcpB transcript levels normalized to this reference always exhibited the lowest values. Between the 2 species examined, EFL-normalized FcpB mRNA fluctuated in a similar range of 0.01 to 5 mol (mol EFL mRNA)⁻¹. However, TBP-normalized FcpB mRNA in C. affinis fluctuated in a wider range and possessed a higher daily mean (Fig. 2B and D).

Fig 2.

Relative gene transcript levels under a light-dark cycle in Skeletonema costatum and Chaetoceros affinis. (A, C) EFL and TBP transcript levels normalized to 18S rRNA in S. costatum and C. affinis, respectively. (B, D) FcpB transcript levels normalized to TBP mRNA, EFL mRNA, and 18S rRNA in S. costatum and C. affinis. Error bars indicate 1 standard deviation of Q-RT-PCR tube replicates. Data points without an error bar mean that the error bar was smaller than the symbol. The filled horizontal bars indicate dark periods.

EFL and TBP sequence diversities in the ECS.

With the use of degenerate primers (Table 1), 32 gene fragments homologous to diatom EFL genes were obtained from genomic DNA extracted from ECS samples. The deduced amino acid sequences of these fragments were aligned with known algal EFL peptides, including 12 partial sequences from GenBank and 3 from diatom cultures in this study, to determine their taxonomic association. A phylogenetic analysis revealed that diatom EFL sequences formed a distinct branch apart from sequences of other algal groups with significant bootstrap support (Fig. 3). Within the diatom branch, the amino acid sequences of EFL shared a mean identity of 85%, but the identity between diatoms and other algae was as low as 46%. Among the 32 sequences from the ECS, 66% were clustered with known sequences from the genera of Chaetoceros, Skeletonema, Pseudonitzschia, and Thalassionema with >60% bootstrap values (Fig. 3). In the case of TBP sequences, we obtained 29 homologous gene fragments from the ECS. Diatom TBP sequences are rare in GenBank, but an alignment of ECS sequences with those from plants, green algae, and diatoms clearly indicated that all 29 sequences were placed in a diatom branch with high bootstrap values (Fig. 4). The mean identity shared by diatom TBP amino acid sequences was 83%, and the identity between diatom and nondiatom groups was 65%. In the 29 ECS partial sequences, 8 were significantly clustered with TBP sequences from 2 cultivated diatoms, S. costatum and C. affinis (Fig. 4). Other sequences also formed distinct clades, but their phylogenetic associations were unclear.

Fig 3.

Neighbor-joining phylogenetic tree of elongation factor-like protein (EFL) amino acid sequences obtained from environmental DNA samples in the East China Sea. EFL sequences obtained in this study are shown in boldface. Numbers at the nodes are the bootstrap values based on 1,000 resamplings, and only values >60% are shown. GenBank accession numbers are shown in parentheses.

Fig 4.

Neighbor-joining phylogenetic tree of TATA box-binding protein (TBP) amino acid sequences obtained from environmental DNA samples in the East China Sea. The TBP sequences obtained in this study are shown in boldface. Numbers at the nodes are bootstrap values based on 1,000 resamplings, and only values >60% are shown. GenBank accession numbers are shown in parentheses. Ske., Skeletonema; Cha., Chaetoceros.

Specificity and amplification efficiency of the group-specific primers.

Group-specific primers were first designed for Chaetoceros and Skeletonema for their high abundances in the ECS (Table 1). Next, we tested the specificity of these primers using genomic DNA from 8 cultivated diatoms. Primers designed for Chaetoceros spp. generated strong amplification signals for EFL, TBP, and Fcp genes from C. affinis but failed to generate PCR products from the other 7 species, indicating that primers were highly specific to Chaetoceros (Fig. 5A). Group-specific primers designed for Skeletonema spp. also performed well with strong amplification in S. costatum. However, some minor (10%) amplification appeared in T. pseudonana and Thalassionema sp. (Fig. 5B).

Fig 5.

Cross-reactivity test of group-specific primers using genomic DNA from 8 cultivated diatoms. Diatom cultures examined were Chaetoceros affinis (Ca), Ditylum brightwellii (Db), Phaeodactylum tricornutum (Pt), Skeletonema costatum (Sc), Thalassionema sp. (Tn), Thalassiosira oceanica (To), T. pseudonana (Tp), and T. weissflogii (Tw). The primer specificity of a diatom was the normalized apparent amount of the target gene with the amount in the target species set to 1.

The amplification efficiencies of these primer sets were evaluated using serially diluted genomic DNA extracted from C. affinis and S. costatum. Results indicated that these primers performed superbly in unialgal cultures, with high amplification efficiencies (E%) of 93.4% to 104% and good linearity (r² > 0.99) (Table 2). Somewhat different results were obtained when the same test was performed again using environmental genomic DNA as the PCR template. The EFL primer sets for Chaetoceros and Skeletonema generated amplification efficiencies of 85.7% and 107%, respectively (Table 2). Although less than perfect, these values were well within the acceptable range for Q-PCR (E% = 100% ± 20%) (37). In contrast, TBP primer sets for both genera generated standard curves with abnormal amplification efficiencies and low linearity (Table 2), which were unsuitable for Q-PCR in environmental samples. In addition, the FcpB primer sets for Skeletonema performed well, but the primer set for Chaetoceros generated high amplification efficiency of 140% (Table 2).

EFL and FcpB transcript levels in the ECS.

Skeletonema group-specific EFL and FcpB transcript abundances were determined in environmental samples, using Q-RT-PCR. With the absolute quantification method, EFL transcript levels of Skeletonema were relatively high during the first half of the sampling period (1200 h on 18 August to 1800 h on 19 August). The range of variation was between 10⁴ and 10⁵ copies (μg total RNA)⁻¹. Later, it declined to around 10³ copies (μg total RNA)⁻¹ between 0600 and 1400 h on 19 August (Fig. 6A). EFL transcript levels of the Chaetoceros group were also high on 18 August and then gradually declined to around 10⁴ copies (μg total RNA)⁻¹ the next day (Fig. 6C). FcpB mRNA levels of the 2 diatom groups revealed 2 peaks, one at noon on 18 August and the other at 0600 h on 19 August (Fig. 6A and C). After using EFL as the reference gene, the transcript levels of FcpB showed a clear diel rhythm with a maximum value after sunset and a minimum value at noon (Fig. 6B and D). Such patterns were similar to those observed in the culture experiments (Fig. 2B and D).

Fig 6.

Group-specific EFL and FcpB transcript levels in the East China Sea detected by Q-RT-PCR. (A, B) Transcript levels obtained using Skeletonema group-specific primers; (C, D) transcript levels obtained using Chaetoceros group-specific primers. Transcript levels were calculated using either the absolute quantification method (A, C) or the relative quantification method (B, D). Relative transcript levels were calculated according to equation 1. Solid bars on the x axis indicate the night period from sunset on 18 August to sunrise on 19 August 2009. Data points without an error bar mean that the error bar was smaller than the symbol.

DISCUSSION

Transcript levels of TBP and EFL were relatively stable under various test conditions including growth stages, light-dark cycle phases, and nutrient stresses (Fig. 1 and 2). Both genes should be good references for normalizing gene expression in culture studies (Fig. 1 and 2). In comparison, TBP expression possessed a 1.33-fold variation during the light-dark cycle and a day-to-day variation of 1.60-fold under various nutrient conditions (Fig. 1D and 2A), which was more stable than EFL expression, with a light-dark cycle variation of 1.59-fold and a day-to-day variation of 2.86-fold (Fig. 1C and 2A). However, group-specific primers for TBP did not perform well for natural phytoplankton, and field applications had to temporarily be suspended (Table 2). The TBP primer sets generated weak amplification signals approaching our detection limit in ECS samples. Part of the reason may have been that the designed primers failed to detect TBP mRNA in natural assemblages. A more-complete sequence database would allow us to design better TBP primers in the future.

Reference genes commonly used in pure cultures, such as 18S rRNA and rbcL, are not suitable for applications in natural environments. Their sequences are so conserved that primers designed for a specific taxonomic group often amplify homologous genes from other phylogenetically unrelated groups (10, 15). This bias would lead to an underestimation of the relative mRNA levels of a target gene and make between-sample comparisons difficult. For this reason, EFL and TBP homologs with less-conserved sequences would be better choices as reference genes. In our cross-activity tests, EFL and TBP primers for C. affinis produced encouraging results in detecting signals only in the target diatom (Fig. 5A). The primers for Skeletonema also detected strong signals from the target species, but a weak apparent detection of 10% was obtained from T. pseudonana as well. This result is not surprising, since Skeletonema is phylogenetically close to several species belonging to the genus Thalassiosira (1). The fact that EFL and TBP mRNAs in other Thalassiosira spp. were not detected suggests that these primer sets are still useful, provided that T. pseudonana is not a dominant species in the plankton samples.

Normalization using transcript abundances of group-specific reference genes is an appropriate strategy to quantify gene expressions in natural phytoplankton samples. In our experiment using ECS samples, the diel variation in FcpB mRNA levels could not properly be revealed on the basis of the per unit mass of total RNA (Fig. 6A and C). Fluctuations in EFL mRNA suggested that the biomass levels of Skeletonema and Chaetoceros varied during the sampling period, probably as a result of water movements. The use of EFL as the reference gene successfully restored the correct patterns of FcpB mRNA levels in both diatom species, with peaks and valleys appearing at similar time points as in laboratory cultures (Fig. 6B and D). This diel expression pattern observed in the sea was consistent with those of typical fcp genes in diatom cultures (22, 27, 28, 32). This is our first instance of recording a clear diel pattern of a functional gene in the sea for specific diatoms.

Gene ratios (DNA ratios) between EFL and FcpB are another piece of evidence to support the applicability of primers designed for Skeletonema. If signals detected in the ECS were truly from the genome of Skeletonema, then the DNA copy ratio of EFL and FcpB genes in the ECS should be similar to that of cultured Skeletonema. Based on the standard curves of EFL and FcpB, the FcpB-to-EFL DNA copy ratios were around 28 for the laboratory culture and 18 for the ECS sample (Table 2). These gene ratios were close to the expected ratio of around 30 in the complete genome of T. pseudonana (16, 38), which indicated that the specificities of EFL and FcpB primer sets were approximately equal for Skeletonema in ECS samples. In addition, diel variations in Skeletonema FcpB expression were 0.02 to 16.6 mol (mol EFL)⁻¹ in ECS samples (Fig. 6B), which were also consistent with the range of 0.04 to 3.70 mol (mol EFL)⁻¹ observed in cultures (Fig. 2B).

In contrast, the performance of the Chaetoceros-specific EFL primer set suggested a broader cross-reactivity than expected. The FcpB-to-EFL DNA ratio was 39.4 in the culture of C. affinis but was <0.01 in an ECS sample (Table 2). In ECS samples, EFL mRNA levels of Chaetoceros were 1 to 2 orders of magnitude higher than those of Skeletonema (Fig. 6A and C), but this result was not supported by cell counts, implying that the EFL primer set could detect EFL mRNA in a wider range of diatom species. Surprisingly, EFL-normalized Chaetoceros FcpB expression in the ECS showed a diel pattern similar to the one observed in the culture of C. affinis (Fig. 2D and 6D). It is likely that although diatom biomass changed during our 24-h sampling period, the species composition remained relatively constant, and mRNA detected by the Chaetoceros-specific EFL primer set still served as a good reference for FcpB expression (Fig. 6D). In the future, the specificity of the Chaetoceros EFL primer set should be further improved by a more-complete investigation of EFL diversity in the sea. In addition, a better interpretation of measured mRNA levels can be achieved by conducting induction/inhibition experiments on deck to obtain the relative position of the in situ sample between the maximum and minimum gene expressions (18, 23).